Learning
that grows
with your child.
Carlo listens while your child reads aloud, corrects at the right moment, and adapts every session to what they actually know.
Personalised. Patient. Proven.
Carlo teaches reading the way the evidence says it should be taught — structured phonics, one sound at a time, with a tutor that waits before it corrects.
It listens properly
Your child reads out loud. Carlo hears the miscue, waits three seconds, then helps at the sound — not the word.
It builds confidence
No buzzers, no red crosses, no score. One specific thing done well, named out loud at the end of every session.
It shows its work
Open the parent app and see exactly why Carlo chose today's lesson, and how far your child has moved.
It's a safe place
Closed to the lesson, no training on your child's data, and a daily cap you set. Safety is the architecture.
Eight minutes,
and almost nothing
on the screen.
Carlo is voice-first. The text your child reads is checked, letter by letter, to contain only the sounds they have already been taught — so a session is reading, not guessing.
- The word being read takes a soft underline. What's coming next stays faint, so the line never feels like a test.
- Nothing moves. No waveform, no animated mouth, no bouncing mascot pulling their eyes off the word.
- Three dots, and that's the session. No streaks, no "one more level". Carlo ends it when the plan is done.
She had a fish in a red dish.
The read-aloud loop. Everything here is also spoken — a session works with the tablet face-down.
A good tutor waits
before they help.
Your child reads fish as "fosh". Carlo says nothing for three seconds — because a child who fixes it themselves has learned something a correction cannot teach. Only then does it step in, at the vowel, in the same warm voice it uses for everything else.
- No red, no cross, no buzzer. A miscue is drawn in the same soft pink as Carlo's own voice.
- Correction at the sound, not the word: "look at the middle sound", then blend it back together.
- The wait is invisible. No countdown ring — a visible timer turns thinking time into pressure.
She had a fish in a red dish.
fishNearly. Look at the middle sound.
Try it again: f — i — sh.
Three seconds of silence is the single most important setting in the product.
Four minutes of listening
Carlo hears your child read once, and turns it into a starting point: which sounds are secure, which are not.
Eight minutes a day
One new sound, then reading aloud on text built from what they already know. Twice a day if they want it.
Growth you can measure
Every six weeks, the same four-minute measure — so progress is a number against age norms, not a feeling.
You paid for a tutor.
You get to check its work.
The same iPad, a different room. Where your child's screen is nearly empty, yours is dense on purpose: what happened this morning, why Carlo taught it, and whether any of it is working.
- Why this lesson. A plain-language paragraph and the numbers behind it — never a naked score.
- Every session's transcript, always, with nothing hidden behind a settings menu.
- Minutes are reported, never celebrated. The headline number is what your child learned per minute of their time.
Léa read for nine minutes this morning.
Session 34 · 07:42–07:51 · full transcript available
Carlo ended the session because the plan was finished, not because the cap was reached.
Parent app, Today. Illustrative data from a 12-week pilot profile.
Safety isn't a policy page.
It's the architecture.
Carlo is closed to the lesson. It is not a friend, not a chatbot, and not somewhere your child can wander.
Anything off-curriculum is caught before it reaches the model and answered with a fixed line.
Every redirect is logged and visible to you.
No training on your child's data. Ever. Raw audio is off by default; we keep a first name, an age band and a skill map.
Export it or delete it in one tap.
Carlo never asks about your home, never offers to keep anything from you, and never presents itself as a substitute for an adult.
No social layer. No other users. Ever.
If your child says anything suggesting they are unsafe, the lesson stops and you are told immediately, with their exact words and local resources.
We built that path before we built the lesson planner.
“We would rather show you the measure than a five-star review.”
Carlo is pre-launch and has no users yet, so there is nothing honest to quote. What we can commit to is the measurement: the same three validated reading measures at baseline and every six weeks, and a pre-registered study with a waitlist control, published whether or not it flatters us.
Validated measures: letter-sound fluency, nonsense-word decoding, and oral reading fluency.
Of your child's time per checkpoint. Every six weeks, on text they have never seen.
Skills in the graph Carlo teaches against, authored with a literacy specialist and versioned like code.
Unverified words. Every passage is either from our vetted corpus or checked, letter by letter, against the sounds your child has been taught.
Fourteen days, one full checkpoint, then you decide.
The trial includes a baseline measure, so you can see where your child actually is before you pay anything.
€20 a month after the trial · cancel in one tap · iPad, iOS 16 and later